pragmatic selection of larval mosquito diets for insectary rearing … · 2019. 8. 19. · 1...
TRANSCRIPT
-
1 Pragmatic selection of larval mosquito diets for
2 insectary rearing of Anopheles gambiae and
3 Aedes aegypti
4
5 Short title: Larval mosquito diets for An. gambiae and Ae. aegypti
6
7 Mark Q. Benedict1 *, Catherine M. Hunt1, Michael G. Vella2, Kasandra M. Gonzalez2, Ellen M.
8 Dotson1, C. Matilda Collins3
9 1 Centers for Disease Control and Prevention, Division of Parasitic Diseases and Malaria,
10 Entomology Branch, Atlanta, Georgia, United States of America
11 2 Frontier Scientific Services, Newark, Delaware, United States of America
12 3 Centre for Environmental Policy, Imperial College London, United Kingdom
13 * Corresponding author
14 Email: [email protected]
15
made available for use under a CC0 license. certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also
The copyright holder for this preprint (which was notthis version posted August 19, 2019. ; https://doi.org/10.1101/740142doi: bioRxiv preprint
https://doi.org/10.1101/740142
-
16 Abstract
17 Larval mosquitoes are aquatic omnivorous scavengers which scrape food from submerged
18 surfaces and collect suspended food particles with their mouth brushes. The composition of
19 diets that have been used in insectaries varies widely though necessarily provides sufficient
20 nutrition to allow colonies to be maintained. Issues such as cost, availability and experience
21 influence which diet is selected. One component of larval diets, essential fatty acids, appears to
22 be necessary for normal flight though deficiencies may not be evident in laboratory cages and
23 are likely more important when mosquitoes are reared for release into the field in e.g. mark-
24 release-recapture and genetic control activities.
25 In this study, four diets were compared for rearing Anopheles gambiae and Aedes aegypti, all of
26 which provide these essential fatty acids. Two diets were custom formulations specifically
27 designed for mosquitoes (Damiens) and two were commercially available fish foods: Doctors
28 Foster and Smith Koi Staple Diet and TetraMin Plus Flakes. Development rate, survival, dry
29 weight and adult longevity of mosquitoes reared with these four diets were measured. The
30 method of presentation of one diet, Koi pellets, was additionally fed in two forms, pellets or a
31 slurry, to determine any effect of food presentation on survival and development rate.
32 While various criteria might be selected to choose ‘the best’ food, the readily-available Koi
33 pellets resulted in development rates and adult longevity equal to the other diets, high survival
34 to the adult stage and, additionally, this is available at low cost.
made available for use under a CC0 license. certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also
The copyright holder for this preprint (which was notthis version posted August 19, 2019. ; https://doi.org/10.1101/740142doi: bioRxiv preprint
https://doi.org/10.1101/740142
-
36 Introduction
37 Larval mosquitoes are omnivorous opportunistic aquatic feeders which collect and swallow
38 small particles, can chew larger particles and can scrape food off of submerged surfaces [1].
39 Laboratory culture of mosquitoes seldom attempts to replicate natural diets but usually consists
40 of a readily available material that experience has proven to allow consistent rearing. These
41 generally fall into two classes: simple mixtures of ingredients such as yeast and liver powder
42 that are formulated by the users or commercial formulations of complex composition including
43 foods such as fish-food flakes [2] or pellets [3], hog supplement [4], cereals and less commonly,
44 maize pollen and algae [5]. Many unusual ingredients such as guinea pig feces and hay infusion
45 are cited by Gerberg [6]. The diversity of ‘successful’ larval foods demonstrates that for many
46 purposes, there are numerous choices.
47 Commercially manufactured diets provide the advantages that the researcher does not need to
48 formulate the diet, can rely on the quality control measures employed by the manufacturer and
49 often, ready availability. Some disadvantages of complex commercial foods are that the
50 researcher has no control over the specific components of the diet which may change without
51 notice and that obtaining the same diet locally in different countries may be difficult.
52 Considerations for choosing a food are simple: It must promote the routine rearing of good
53 quality mosquitoes (however that is defined), be readily available, consistent in quality and
54 preferably inexpensive. A less apparent and seldom considered advantage is to choose a diet
55 that numerous laboratories can use to give some assurance of comparable results. If flight
56 testing, mating assays or release activities will be performed, it is necessary to provide essential
57 fatty acids [7].
58 This study determined, among four candidate larval diets for two frequently-reared disease
59 vector mosquitos, Anopheles gambiae Giles (Diptera: Culicidae) and Aedes aegypti Linnaeus
made available for use under a CC0 license. certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also
The copyright holder for this preprint (which was notthis version posted August 19, 2019. ; https://doi.org/10.1101/740142doi: bioRxiv preprint
https://doi.org/10.1101/740142
-
60 (Diptera: Culicidae), which diet and feeding level resulted in the optimal performance for several
61 important life history traits such immature growth rate, survival, size and adult longevity. One of
62 these diets, TetraMin flakes, is widely used for both Anopheles and Aedes spp. We make a
63 recommendation for selection among these diets which also considers cost and availability.
64 Materials and Methods
65 Diet preparation
66 Four diets were prepared for comparisons; two of these were custom formulations of a diet
67 specifically designed for mosquitoes [8]. This diet consists of a 2:2:1 ratio (by weight) of bovine
68 liver powder, tuna meal and Vanderzant vitamin mix. One formulation was prepared at CDC in
69 Atlanta, GA using ‘Now’ brand liver powder (Bloomingdale, IL USA), tuna meal (AA Baits, Rock
70 Ferry, Birkenhead, UK) and Vanderzant vitamin mix (Bio-Serv, Flemington, NJ, USA). Large
71 particles were removed from the tuna meal and liver powder using a (600 ) standard sieve.
72 Clumps of vitamin mix were broken up manually but no further sieving was done because the
73 mix is soluble and the particle size allowed even mixing.
74 The other formulation of the Damiens diet was prepared by Frontier Scientific Services (Newark,
75 DE USA) using defatted, desiccated liver powder (product no. 1320; Frontier Scientific
76 Services), Vanderzant vitamin mix (product no. F8045, Frontier Scientific Services) and the
77 same lot of tuna meal as was used at CDC. In order to ensure particle size was small enough
78 for consumption by developing larvae, a milling and screening procedure was employed. A
79 significant source of oversized particulates was the tuna meal. Most large particles were
80 identified as scale and bone remnants from the manufacturers processing of the meal. The tuna
81 meal was processed in a top-feeding hammer mill (The Fitzpatrick Co., Toronto, Canada) with a
82 60-80 (177 mesh particle excluding screen. To ensure particles were milled to specification
made available for use under a CC0 license. certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also
The copyright holder for this preprint (which was notthis version posted August 19, 2019. ; https://doi.org/10.1101/740142doi: bioRxiv preprint
https://doi.org/10.1101/740142
-
83 without complete exclusion of meal components, the material was passed through the hammer-
84 mill twice. The milled tuna meal was then mixed with the remaining ingredients in a bench-top
85 ‘Kitchen Aid’ bread mixer for 20 minutes. After mixing was complete, the final diet was hand
86 sifted through a 60 mesh (177 screen to eliminate any remaining oversized particulates.
87 The other two diets were commercially available fish foods: Doctors Foster and Smith Koi
88 Staple Diet (Rhinelander, WI USA) and TetraMin Plus Flakes (Tetra GmbH, Melle, Germany).
89 For fair comparison, both fish foods were ground to a similar size to the custom diets. Koi pellets
90 were ground in a Miracle Model MR-300 Electric Grain and Flour Mill (Danbury, CT USA)
91 followed by sieving in a 600 standard sieve and saving the particles that passed. The
92 TetraMin was ground in a Black and Decker ‘SmartGrind’ coffee grinder (Beachwood, OH USA)
93 after which it easily passed through a 600 sieve. These diet types will be identified as CDC,
94 Frontier, Koi, and TetraMin respectively.
95 The ground diets were mixed at 0.4, 0.8, 1.6 and 3.2 % w/v in type II water and stored in ca. 13
96 ml aliquots and frozen at -20C where they remained until being thawed in warm water
97 immediately before feedings. When 4 ml of the slurry was fed, these concentrations result in
98 feeding rates (levels) of 8, 16, 32 and 64 mg diet / dish / day. Hereafter we will usually refer to
99 the levels simply as e.g. 32 mg.
100 Mosquitoes
101 Anopheles gambiae mosquitoes were the ‘G3’ strain (MRA-112) obtained from the Malaria
102 Research and Reference Resource Center (MR4, BEI Resources, Manassas VA USA). Aedes
103 aegypti were the ‘New Orleans’ strain (NR-49160), also obtained from the MR4 and were in the
104 F16-F18 generations during experiments. A standard rearing water was made of 0.3 g of pond
105 salts (API, McLean, VA USA) per liter of type II purified water. Anopheles gambiae eggs were
made available for use under a CC0 license. certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also
The copyright holder for this preprint (which was notthis version posted August 19, 2019. ; https://doi.org/10.1101/740142doi: bioRxiv preprint
https://doi.org/10.1101/740142
-
106 collected, held overnight on damp filter paper and placed in trays on the day of hatching. Aedes
107 aegypti eggs were hatched by placing egg papers in water under vacuum for 30 min. Hatching
108 embryos of both species were placed in 500 ml of water containing 5 intact Koi pellets for one
109 day before counting 80 larvae into 150 ml polystyrene Petri dishes (Item no. Z717231, Sigma-
110 Aldrich, St. Louis, MO USA).
111 Trial design
112 An orthogonal design was used; three dishes (replicates) for each of the four diets at all four
113 levels were established for both species. For these mosquitoes, it is not possible to determine
114 sex at the first larval instar and it was assumed that random aliquots would deliver a
115 representative sex ratio. Before counting larvae into Petri dishes, the empty dishes were
116 weighed to a tenth of a gram on a triple-beam balance (700/800 Series, Ohaus, Parsippany, NJ
117 USA) and labeled with their weight. On the day after hatching, 80 larvae were counted into the
118 dishes and rearing water was added until the net weight was 96 g. Then 4 ml of food was added
119 for an approximate total volume of 100 ml. The concentrations were selected to bracket a range
120 shown to allow maximal survival and development rate with Anopheles arabiensis [9]. Additional
121 diet was added on alternate days, prior to which the dishes were weighed and water was
122 removed (ca. 2-3 ml), to return the net weight to 96 g before 4 ml of diet slurry was added to
123 maintain an approximate total volume of 100 ml. Mosquitoes were reared in an environmental
124 room set at 27C and 70% relative humidity with a 12:12 light:dark cycle and 30 minutes of
125 dawn and dusk.
126 In this main experiment, in which all diets and levels were tested, pupae were counted and
127 collected daily, their sex determined and the pupae were then placed in individual plastic tubes
128 for eclosion. Tubes were checked for adults daily with up to five randomly selected adults from
129 each day of eclosion and sex being killed for dry weight measurements. Immature stage trials
made available for use under a CC0 license. certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also
The copyright holder for this preprint (which was notthis version posted August 19, 2019. ; https://doi.org/10.1101/740142doi: bioRxiv preprint
https://doi.org/10.1101/740142
-
130 were generally terminated when there were no more larvae present except as noted for the 8
131 mg diet level with An. gambiae where observations of larval duration were terminated based on
132 a pragmatic decision on days 12 and 14 (Table 1).
133 Anopheles gambiae and Ae. aegypti differ in many characteristics including body size and
134 rearing tractability. Because of this, the two species have been analyzed separately. There are
135 also known differences in outcomes by sex within species and, where appropriate, parameter
136 estimates for each sex are calculated independently. Statistical analyses were performed using
137 R version 3.5.1 “Feather Spray” [10]
138 Inter-trial comparison of water temperature
139 Due to logistical limitations, it was not possible to perform all experiments concurrently. As a
140 result, five sequential trials in the same chambers contributed to the experiment overall. The
141 critical variable of water temperature was measured in three arbitrary dishes every two or three
142 days in the morning using a Sper Scientific Model 800005 thermocouple thermometer
143 (Scottsdale, AZ USA) equipped with a K type probe and overall means were compared using
144 Analysis of Variance.
145 Sex ratio
146 The sex of pupae arising from each treatment combination was observed and the ratio
147 estimated. Chi-square tests were used to determine whether the male:female ratio varied with
148 treatment or species.
149 Survival to eclosion
150 To determine the effect of the different diets and levels fed on the number of adults that eclosed,
151 Poisson-family generalized linear models (GLMs) using the diet level, diet type and their
made available for use under a CC0 license. certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also
The copyright holder for this preprint (which was notthis version posted August 19, 2019. ; https://doi.org/10.1101/740142doi: bioRxiv preprint
https://doi.org/10.1101/740142
-
152 interaction were fit to the data. Model simplification by deletion tests used F-tests to estimate
153 influential effects as appropriate to the over-dispersion of these count data.
154 Proportion of pupae eclosing
155 The data that were analyzed resulted from the counts of the number of pupae that formed and
156 eclosion data. A weighted response variable that bound the number of pupae eclosing and the
157 number of pupae that did not was created. Binomial-family GLMs using the dose, food type
158 (both categorical, four levels) and their interactions were fit to these data. Model simplification
159 by deletion used F-tests to estimate important effects as appropriate to the over-dispersion
160 evident in the weighted proportion data.
161 Larva developmental rate
162 The number of days taken to complete larval development to pupation was analyzed to
163 determine effects on development rates. As the number of days to eclosion was an integer
164 value, chi-squared tests were used to estimate the influence on this time of interactions between
165 diet type and the amount of food provided as well as these as single effects. The relative
166 contribution of each factor is then reflected in the test statistic values.
167 Adult longevity estimation
168 The 32 mg diet level was chosen for assessing the influence of diet type on adult longevity
169 based on observed rapid development rate and high survival across all diet types reported in
170 the Results section. Pupae arising from these dishes were placed in aluminum-frame cages [11]
171 which were covered with one or two layers (in the case of Ae. aegypti) of gauze and provided
172 sugar water (10% w/v food grade sucrose, 0.1% w/v methylparaben in type II water) which was
173 changed weekly. There were three cages for each diet, each associated with a different larval
made available for use under a CC0 license. certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also
The copyright holder for this preprint (which was notthis version posted August 19, 2019. ; https://doi.org/10.1101/740142doi: bioRxiv preprint
https://doi.org/10.1101/740142
-
174 replicate dish. All longevity measures were made concurrently. Mortality was usually checked
175 daily, though occasionally it was not observed on Saturdays. Kaplan-Meier objects were created
176 as response variables for the survival analyses and a Cox proportional hazards model was used
177 to identify effects of diet type on survival for each species and sex.
178 Measures of dry weight
179 Dry weight of adults was determined for all diet types and levels. After eclosion, adults were
180 transferred to glass scintillation vials, killed by freezing at -20C and dried in a drying oven at
181 60C overnight after which they were removed and the caps sealed until weighing. For each
182 diet/level combination, up to five individuals of each sex from each day of eclosion were
183 weighed using a Sartorius SuperMicro S4 balance (Bohemia, NY USA) when that number was
184 available. Weights are reported in micrograms.
185 The dry weight of mosquitos was a continuous response variable. As previously, diet type, level
186 and mosquito sex were all considered as categorical factors. All main effects and interactions
187 were tested by deletion from the maximal model. Effects that were either non-significant or
188 accounted for less than 1% of the variation in the data were excluded.
189 The influence of pellet vs. slurry
190 One food type, the Koi pellets, was used to estimate any influence of the form of presentation
191 and thus whether it is necessary to grind the food. Koi pellets were weighed on the SuperMicro
192 balance and the average weight and standard deviation of pellets calculated; 52.0 mg (n=14,
193 StDev 8.65). Two pellets (equivalent to 52 mg/dish/day) were fed on alternate days in parallel
194 with the day the 32 mg slurry was given. Larval survival (the number of larvae reaching
195 pupation) and larval duration (the number of days to pupation) were used as the measures for
made available for use under a CC0 license. certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also
The copyright holder for this preprint (which was notthis version posted August 19, 2019. ; https://doi.org/10.1101/740142doi: bioRxiv preprint
https://doi.org/10.1101/740142
-
196 this comparison. Pupae were collected daily in the morning and their sex determined. All dishes
197 were new and there were three tests of each food form for both Ae. aegypti and An. gambiae.
198 Results
199 Inter-trial comparability of water temperature
200 The temperature was consistent among all trials of Ae. aegypti (F=1.03, d.f.= 2,72 p=0.36). The
201 average water temperature was 26.9C (n=75, StDev 0.45). The average temperature of all
202 An. gambiae trials was 27.0°C (n=45, StDev 0.33) but there was a slight, but significant,
203 variation in temperature between the trials (F=7.51, d.f.=1.43, p
-
Table 1: The number of An. gambiae immatures discarded at the end of the trial at the lowest diet level, 8mg.
Diet type Dish Day Discarded
A 47
B 28CDC
C
14
48
A 13
B 7Frontier
C
14
10
A 51
B 52Koi
C
12
47
A 63
B 66TetraMin
C
12
57
217
218 The responses of Ae. aegypti and An. gambiae to different diets and levels shared similarities in
219 pattern but had marked differences in absolute level (Fig 1). Aedes aegypti eclosion varied less
220 as a function of diet type and level and achieved higher numbers than did An. gambiae (for
221 model null deviance see Table 2). There was an interaction between diet type and level on the
222 number of Ae. aegypti males and females; this was largely driven by the two commercial foods,
223 Koi and TetraMin, having higher numbers at the lowest diet level than did the CDC or Frontier
224 diets.
225 Fig 1. The number of Ae. aegypti and An. gambiae female and male adults observed by
226 diet type and level. The dashed horizontal line indicates the expected number of females and
227 males assuming a 1:1 sex ratio and full survival. Error bars are the 95% CI of the mean.
228 Darkening shades of color represent the increasing diet levels of 8, 16, 32 and 64 mg.
made available for use under a CC0 license. certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also
The copyright holder for this preprint (which was notthis version posted August 19, 2019. ; https://doi.org/10.1101/740142doi: bioRxiv preprint
https://doi.org/10.1101/740142
-
Table 2. Statistical summary of the influences on the number of male and female adults
formed with the proportion of the deviance explained (in parentheses) given as an indicator
of effect size (significant effects indicated in bold).
Ae. aegypti An. gambiae
Males Females Males Females
Model null deviance (47 d.f.)
86.77 201.09 689.35 669.89
Diet:Level F=5.11, d.f.= 9,32, p
-
237 Proportion of pupae eclosing
238 We anticipated that the likelihood of pupae that had formed then successfully eclosing might be
239 affected by the food type or level. For both sexes of Ae. aegypti there was an interaction
240 between diet type and level on the number of pupae eclosing to adults (Table 3); this was
241 largely driven by poor eclosion at low levels of the Frontier diet (Fig 2). In all other cases, if the
242 larvae reached the pupa stage, they were highly likely to become an adult.
243 Fig 2. Eclosion of pupae that formed by diet type and level. Error bars represent the 95%
244 CI. Darkening shades of color represent the increasing diet levels of 8, 16, 32 and 64 mg.
Table 3. Model summary statistics estimating the influence of diet type and level on
the number of male and female pupae eclosing to adults with the effect size
(proportion of the deviance explained) indicated in parentheses (significant effects
in bold).
Ae. aegypti An. gambiaeMales Females Males Females
Model null deviance
165.10 155.63 207.64 96.12
Diet:LevelF=3.33,
d.f.=9,32, p = 0.006 (0.35)
F=3.19, d.f.=9,32,
p=0.007 (0.33)
F=1.76, d.f.=8,25, p=0.14 (0.14)
F=1.15, d.f.= 8,26,
p=0.37 (0.14)
Diet F=1.14, d.f.=3,41, p=0.35 (0.07)
F=1.43, d.f.= 3,41, p=0.25,
(0.08)
F=4.31, d.f.=3,33, p=0.011(0.16)
F=4.53, d.f.= 3,34, p=0.009
(0.19)
made available for use under a CC0 license. certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also
The copyright holder for this preprint (which was notthis version posted August 19, 2019. ; https://doi.org/10.1101/740142doi: bioRxiv preprint
https://doi.org/10.1101/740142
-
Level F=2.24 d.f.=3,41, p=0.10, (0.13)
F=2.51, d.f=3,41, p=0.07 (0.14)
F=10.10, d.f.=3,33,
p
-
Ae. aegypti An. gambiae
Females Males Females Males
Interaction
Diet type:Level
2 = 37.46, df = 9,
p < 0.001
2 = 26.31,df = 9,
p < 0.002
2 = 19.28,df = 8,
p < 0.05
2 = 19.19,df = 6,
p < 0.01
Diet type2 = 12.77,
df = 3, p< 0.01
2 = 25.63,df = 3,
p0.05 in all cases). Overall, the median adult lifespan of Aedes aegypti males and females
274 was similar (Table 5). For females, there was no identifiable variation in longevity as a function
made available for use under a CC0 license. certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also
The copyright holder for this preprint (which was notthis version posted August 19, 2019. ; https://doi.org/10.1101/740142doi: bioRxiv preprint
https://doi.org/10.1101/740142
-
275 of diet type (2=4.45, d.f.=3, p=0.22). There was variation in male longevity but CDC and Koi
276 led to longer-lived males (2=12.20, d.f.=3, p=0.007).
277
278 Anopheles gambiae females lived consistently longer than males (Table 5). For females, diet
279 type affected longevity with CDC and Koi leading to longer life (2=9.87, d.f.=3, p=0.02). There
280 was greater variation in male longevity but no diet-related variation was identified (2=5.80,
281 d.f.=3, p=0.12).
282
Table 5. Adult Longevity
Ae. aegypti
Females Males
Diet n Median (95% CI) n Median (95% CI)
CDC 55 51 (46-67) 80 59 (56-67)
Frontier 52 49 (35-67) 90 54 (51-61)
Koi 58 57 (53-63) 82 60 (57-63)
TetraMin 69 50 (31-68) 80 49 (44-53)
An. gambiae
Females Males
Diet n Median (95% CI) n Median (95% CI)
CDC 47 37 (25-39) 55 21 (14-26)
Frontier 57 32 (29-37) 62 29 (27-32)
Koi 53 37 (37-37) 48 20 (15-30)
TetraMin 66 30 (26-37) 56 24 (18-29)
made available for use under a CC0 license. certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also
The copyright holder for this preprint (which was notthis version posted August 19, 2019. ; https://doi.org/10.1101/740142doi: bioRxiv preprint
https://doi.org/10.1101/740142
-
283 Dry weight
284 Aedes aegypti dry weight
285 A total of 787 Ae. aegypti females and 880 males were weighed. Aedes aegypti males weigh
286 less than females (p
-
Sex (factor) 1770.50 1,1659
-
Sex:Level 0.58 0.00
Minimal adequate model 35.78 15,770
-
315 female larvae fed pellets delayed pupation by a day (median values: Ae. aegypti 7:6,
316 An. gambiae 9:8 pellet vs. slurry respectively).
317 Discussion
318 In this diet comparison, a range of diets fed at rates ranging from very low to high was
319 compared. This experimental design was chosen to reduce the likelihood that the variation in
320 the proportion of any particular component of diet (protein, fat or carbohydrates) might result in
321 outcomes that do not represent the most favorable levels of diet fed. Because the ratios of
322 protein, carbohydrates and fats differ among diets, a wide-level design is agnostic regarding
323 which is most important for the outcomes tested. This approach is in contrast to Linenberg [3] in
324 which the combined weight of fat and protein – to the exclusion of carbohydrates - was used to
325 determine the amounts of diet provided to larvae for comparisons.
326 The reputation of Ae. aegypti as a robust and physiologically plastic laboratory model for
327 laboratory study was borne out by the high eclosion rates at all diet levels compared to
328 An. gambiae which was very sensitive to level. This trait also makes it a relatively insensitive
329 choice with which to compare diets.
330 These results demonstrated that as far as choosing a diet, TetraMin is the least desirable for
331 An. gambiae because of the sensitivity to diet level that was required for adult production;
332 neither the highest nor lowest doses resulted in adults within what we considered a practical
333 time period. Linenberg et al. [3] also observed that two pellet fish foods performed better than
334 TetraMin flakes though it is not clear whether the specific product was the same as the one we
335 tested.
336 We were surprised that two different formulations of the Damiens diet prepared by CDC and
337 Frontier Scientific Services gave measurably different results. There are two differences which
made available for use under a CC0 license. certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also
The copyright holder for this preprint (which was notthis version posted August 19, 2019. ; https://doi.org/10.1101/740142doi: bioRxiv preprint
https://doi.org/10.1101/740142
-
338 might have contributed. Frontier used defatted liver powder whereas the CDC source did not
339 specify whether it was defatted or not. Secondly, the Frontier team had access to a hammer mill
340 which permitted the tuna meal to be ground more finely – likely contributing a larger amount of
341 indigestible scale and bone to the final formulation of diet resulting in lower concentrations of
342 other tuna parts. The CDC team discarded the larger particles. These two differences may have
343 resulted in a formulation with substantially different nutritional content on a weight basis.
344 Of the diets tested, one can make an evaluation of their performance assuming, somewhat
345 subjectively, that maximal survival rates, longevity and size along with short development times
346 are desirable outcomes (Table 9).
Table 9. A semi-subjective assessment of the salient biological outcomes measured as an
assessment of laboratory use of the four diets tested (advantageous characteristics are highlighted
in green, neutral ones in gray and disadvantageous ones in yellow.)
Ae. aegypti An. gambiae
CDC Frontier Koi TetraMin CDC Frontier Koi TetraMin
Survival to
eclosionNo effect of diet type Highest
Probability of
pupae to
eclose
No effect of diet type Highest
Eclosion
sensitivity to
diet levels
fewer
adults
eclosing
Lowest
made available for use under a CC0 license. certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also
The copyright holder for this preprint (which was notthis version posted August 19, 2019. ; https://doi.org/10.1101/740142doi: bioRxiv preprint
https://doi.org/10.1101/740142
-
at lowest
level
Development
rateLittle difference observed
More
consiste
nt
Dry weight Little difference observed Little consistent advantage
Adult
longevity
> for
males
> for
males
> for
females
> for
females
347
348 The deviation from a 1:1 ratio of females and males that we observed in the New Orleans strain
349 of Ae. aegypti is common among many strains of Ae. aegypti [12]. In contrast, the authors are
350 unaware of any natural strains of An. gambiae that demonstrate sex ratio bias although this has
351 been observed among progeny of crosses between different species of the An. gambiae
352 complex [13].
353 One diet, Koi, was tested to determine whether the method of presentation of the same diet had
354 an effect on the development rate and survival to the pupa stage. Of the other diets that could
355 be fed in either a whole or ground form, only TetraMin is originally in a flake form and similar
356 comparisons are possible. Any of the powders or flakes can be fed either as powder sprinkled
357 on the surface, a practice which is consistent with the ‘surface feeding’ behavior of Anopheles
358 spp. [1].
359 The authors are aware that some laboratories provide the diet as intact pellets or flakes rather
360 than as a slurry. The difference between the total weights of food in our analysis confounds our
361 analysis and arguably, if one provided more pellets, the development rates of females would be
362 the same as when fed slurry. But as far as these analyses can be interpreted, one can conclude
363 that for a given amount of food, increasing the availability in a ground form will increase the
made available for use under a CC0 license. certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also
The copyright holder for this preprint (which was notthis version posted August 19, 2019. ; https://doi.org/10.1101/740142doi: bioRxiv preprint
https://doi.org/10.1101/740142
-
364 development rate. Feeding as a slurry also allows a continuously variable (rather than discrete)
365 amount of food to be delivered though this advantage requires mixing and pipetting slurry vs.
366 simply counting pellets.
367 Our results demonstrate that although the An. gambiae feeding rate in mosquito publications on
368 is often described as ‘ad libitum’, it is almost certain this is not the case. The levels of diet that
369 result in the largest size cause so much mortality that they would not be used. Expressing it
370 another way, larvae will continue eating more food at levels that are not consistent with overall
371 survival of mosquitoes for experiments. In most experiments, the amount of food that is made
372 available always restricts growth below the maximal size possible under true ad libitum
373 conditions.
374 We consider all of the diets tested acceptable in our hands. However, the superior performance
375 and low cost of the Koi food makes it a good choice for most purposes. It can be fed either as a
376 slurry or pellet and is available in large amounts which can be frozen to stockpile the food for
377 future use, a practice that would permit only occasional importation.
378
made available for use under a CC0 license. certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also
The copyright holder for this preprint (which was notthis version posted August 19, 2019. ; https://doi.org/10.1101/740142doi: bioRxiv preprint
https://doi.org/10.1101/740142
-
379 Acknowledgments
380 We appreciate the custom Damiens diet formulation that was generously prepared by MV and
381 KG at Frontier Scientific Services and supplied to the CDC with the understanding that the
382 experimental design and diet comparisons would not be influenced by the potential for
383 commercialization. Frontier kindly formulated the diet and provided it without charge for these
384 comparisons.
385 The following reagents were obtained through the NIH Biodefense and Emerging Infections
386 Research Resources Repository, NIAID, NIH: An. gambiae, strain ‘G3’ (MRA-112) and Ae.
387 aegypti ‘New Orleans’ strain (NR-49160).
388 The findings and conclusions in this report are those of the authors and do not necessarily
389 represent the official position of the Centers for Disease Control and Prevention.
390 Funded byTarget Malaria, a project that receives core funding from the Bill & Melinda
391 Gates Foundation and from the Open Philanthropy Project Fund, an advised fund of Silicon
392 Valley Community Foundation.
393
394
made available for use under a CC0 license. certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also
The copyright holder for this preprint (which was notthis version posted August 19, 2019. ; https://doi.org/10.1101/740142doi: bioRxiv preprint
https://doi.org/10.1101/740142
-
395 References Cited
396 1. Clements AN. The biology of mosquitoes. Volume 1: development, nutrition and
397 reproduction. CABI;1992.
398 2. Carvalho DO, Nimmo D, Naish N, McKemey AR, Gray P, Wilke, ABB, et al. Mass
399 Production of Genetically Modified Aedes aegypti for Field Releases in Brazil. J
400 Visualized Exps. 2013; (), e3579.
401 3. Linenberg I, Christophides GK, Gendrin M. Larval diet affects mosquito development and
402 permissiveness to Plasmodium infection. Sci Rep. 2016;6(1):38230–10.
403 4. Dame DA, Haile DG, Lofgren CS, Bailey DL, Munroe WL. Improved rearing techniques
404 for larval Anopheles albimanus: use of dried mosquito eggs and electric heating tapes.
405 Mosquito News. 1978;38(1):68–74.
406 5. Kivuyo HS, Mbazi PH, Kisika DS, Munga S, Rumisha SF, Urasa FM, et al. Performance
407 of Five Food Regimes on Anopheles gambiae senso stricto Larval Rearing to Adult
408 Emergence in Insectary. PLoS ONE. 2014;9(10):e110671.
409 6. Gerberg EJ. Manual for mosquito rearing and experimental techniques. Am Mosquito
410 Control Assoc Bull. 1979;5:1–124.
411 7. R DH, Kleinjan JE, Asman SM. Eicosapentanoic Acid in Mosquito Tissues: Differences
412 Between Wild and Laboratory-Reared Adults. Environ Entomol. 1988;17(2):172–80.
413 8. Damiens DD, Benedict MQ, Wille M, Gilles JRL. An Inexpensive and Effective Larval Diet
414 for Anopheles arabiensis (Diptera: Culicidae): Eat Like a Horse, a Bird, or a Fish? J Med
415 Entomol. 2012;49(5):1001–11.
made available for use under a CC0 license. certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also
The copyright holder for this preprint (which was notthis version posted August 19, 2019. ; https://doi.org/10.1101/740142doi: bioRxiv preprint
https://doi.org/10.1101/740142
-
416 9. Gilles JRL, Lees RS, Soliban SM, Benedict MQ. Density-dependent effects in
417 experimental larval populations of Anopheles arabiensis (Diptera: Culicidae) can be
418 negative, neutral, or overcompensatory depending on density and diet levels. J Med
419 Entomol. 2011;48(2):296–304.
420 10. R Core Team (2018). R: A language and environment for statistical computing. R
421 Foundation for Statistical Computing, Vienna, Austria. URL https://www.R-project.org/
422 11. Savage KE, Lowe RE. A one-piece aluminum cage designed for adult mosquitoes.
423 Mosquito News. 1971;31(1):111–2.
424 12. Hickey WA, Craig GB. Genetic distortion of sex ratio in a mosquito, Aedes aegypti.
425 Genetics. 1966;53(6):1177–96.
426 13. Davidson G. Anopheles gambiae, a complex of species. Bull WHO. 1964;31:625–34.
made available for use under a CC0 license. certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also
The copyright holder for this preprint (which was notthis version posted August 19, 2019. ; https://doi.org/10.1101/740142doi: bioRxiv preprint
https://doi.org/10.1101/740142
-
made available for use under a CC0 license. certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also
The copyright holder for this preprint (which was notthis version posted August 19, 2019. ; https://doi.org/10.1101/740142doi: bioRxiv preprint
https://doi.org/10.1101/740142
-
made available for use under a CC0 license. certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also
The copyright holder for this preprint (which was notthis version posted August 19, 2019. ; https://doi.org/10.1101/740142doi: bioRxiv preprint
https://doi.org/10.1101/740142
-
made available for use under a CC0 license. certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also
The copyright holder for this preprint (which was notthis version posted August 19, 2019. ; https://doi.org/10.1101/740142doi: bioRxiv preprint
https://doi.org/10.1101/740142
-
made available for use under a CC0 license. certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also
The copyright holder for this preprint (which was notthis version posted August 19, 2019. ; https://doi.org/10.1101/740142doi: bioRxiv preprint
https://doi.org/10.1101/740142
-
made available for use under a CC0 license. certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also
The copyright holder for this preprint (which was notthis version posted August 19, 2019. ; https://doi.org/10.1101/740142doi: bioRxiv preprint
https://doi.org/10.1101/740142
-
made available for use under a CC0 license. certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also
The copyright holder for this preprint (which was notthis version posted August 19, 2019. ; https://doi.org/10.1101/740142doi: bioRxiv preprint
https://doi.org/10.1101/740142
-
made available for use under a CC0 license. certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also
The copyright holder for this preprint (which was notthis version posted August 19, 2019. ; https://doi.org/10.1101/740142doi: bioRxiv preprint
https://doi.org/10.1101/740142